PoC: Faster HANA streaming

[BobdenOs]

This PoC explorers a possible way to avoid certain network protocol deficiencies. To unlock greater single connection and overall throughput capabilities for the @cap-js/hana service.

The current performance measurements can be found here.

For the measurements of this PR node@24 was used as this comes with optimizations that also improve throughput.

mode	hana main	hana faster	postgres main
sync	150-250 rows/ms	200-300 rows/ms	100-120 rows/ms
async	350-450 rows/ms	1000-1200 rows/ms	730-750 rows/ms

In the case of postgres the same approach observed a large reduction in performance (~50%), caused by materialization inside postgres causing memory troubles. By just upgrading to node@24 the postgres throughput went up by ~50%. For HANA once the network package management is avoided the single connection performance becomes more stable. The major impact is on the throughput as the network capacity is utilized to its fullest extend. Enabling the capability to download 1million books in a single second. With a peak application memory of ~50mb during the whole process.